Temperature gradients across the dies of conventional high performance very large scale integration (VLSI) components, such as a microprocessor, can adversely affect component performance. For example, a temperature variation between two clock driver circuits within a microprocessor often results in a skew in the system clock of the microprocessor. Moreover, the die of the microprocessor may reach an unacceptable temperature that causes the microprocessor to malfunction or stop functioning.
A diode is often placed in a die of the microprocessor to provide a die temperature indication. This diode is driven with a fixed amount of current, and the corresponding voltage drop across the diode provides an indication of the microprocessor temperature. Unfortunately, the diode provides a temperature reading that can be accurate to about only ±10° C., which is often not accurate enough to provide an early indication of a temperature abnormality. Moreover, a single diode is typically utilized to monitor the die temperature of the entire microprocessor. Given the size and complexity of current and future microprocessors, it is impossible to determine and monitor a temperature gradient across the microprocessor using only a single diode positioned at a location in the die of a microprocessor. As such, substantial variations in temperature across the die of the microprocessor can go undetected. Consequently, any indication that a thermal gradient problem exists in a portion of the integrated circuit go undetected.